How betting a UFC fight actually works

The simplest UFC bet is the moneyline — pick the winner. US books quote it in American odds: a negative number is the favorite, a positive number the underdog.

Every price implies a probability. For favorites: odds ÷ (odds + 100). For dogs: 100 ÷ (odds + 100). Holloway at -210 implies 210/310 ≈ 68%; McGregor at +195 implies 100/295 ≈ 34%. Notice those sum to 102%, not 100% — that extra 2% is the next section.

The bookmaker's margin — the vig — is baked into every price. Remove it proportionally and you get the fair price: what the odds would be if the book kept nothing. Serious bettors compare everything against fair prices, because comparing a prediction to a vigged price flatters the book, not you.

On this site, every edge is measured fair-vs-fair— our model's probability against the no-vig fair implied by the sharpest books. The methodology page has the details.

Beyond the winner, books price how a fight ends — KO/TKO, submission, or decision — and when:

• Round totals (e.g. Over/Under 2.5 rounds): does the fight last past the midpoint of round 3?
• Inside the distance: does it end before the final bell, either way?
• Method props: a specific fighter by a specific method, sometimes in a specific round.

These markets are thinner than the moneyline — fewer bettors, wider margins, slower prices. That cuts both ways: prices are worse on average, but mistakes live longer.

A price isn't one number, it's a history. Books post an opener, money arrives, and the price drifts until the close— the last price before the cage door shuts. The close is the market's most informed number, because it has absorbed every bet.

Different books also hold different prices at the same moment. Line shopping — holding accounts at several books and taking the best available price on the bet you already wanted — is the single easiest improvement any bettor can make. It costs nothing and compounds forever.

An edge is a disagreement you can quantify: you believe an outcome is more likely than the fair price says. We express it in probability points (pp) — model 55% vs fair 48% is a +7pp edge. Small numbers matter here; a durable +2–3pp is a serious advantage, and anyone advertising +20pp edges every card is selling something.

One honest caveat that most sites skip: a model's raw output is usually overconfident. Ours is calibrated — corrected against its own historical misses — before any edge is shown.

Bettors measure results in units — one unit is one standard stake, usually 1–2% of bankroll. Units make records comparable and keep any single fight from mattering too much, because even great bettors lose constantly: at a 56% win rate, losing streaks of seven or eight are expected over a year.

The discipline that survives variance is boring on purpose: flat stakes, fixed rules, no chasing, no doubling after losses. If a losing streak would change your stake size, the stake was too big.

Profit is noisy; closing line value is not. CLV asks: was the price you locked better than the closing price? Beat the close consistently and profit follows; lose to the close consistently and no hot streak will save you. It's the strongest single signal that a bettor has edge rather than luck — which is why our track record reports it next to ROI.

Everything free here is a view of one model: the win bars (who), the method grid and finish forecast (how and when), and the model edges board — the places our number disagrees with the books enough to become an official play. Official plays are hash-committed before every card and graded publicly after it, win or lose; the methodology page explains how to verify that yourself.